[1]毕卉,杨冠羽,唐慧,等.基于改进主动形状模型的前列腺超声图像分割算法[J].东南大学学报(自然科学版),2017,47(5):879-883.[doi:10.3969/j.issn.1001-0505.2017.05.007]
　Bi Hui,Yang Guanyu,Tang Hui,et al.Prostate ultrasound image segmentation algorithm based on improved active shape model[J].Journal of Southeast University (Natural Science Edition),2017,47(5):879-883.[doi:10.3969/j.issn.1001-0505.2017.05.007]
点击复制

基于改进主动形状模型的前列腺超声图像分割算法()
分享到：

《东南大学学报(自然科学版)》[ISSN:1001-0505/CN:32-1178/N]

卷:
47
期数:
2017年第5期
页码:
879-883
栏目:
计算机科学与工程
出版日期:
2017-09-20

文章信息/Info

Title:
Prostate ultrasound image segmentation algorithm based on improved active shape model
作者:
毕卉1; 杨冠羽1; 2; 3; 唐慧1; 2; 3; 舒华忠1; 2; 3
1东南大学计算机科学与工程学院, 南京210096; 2东南大学计算机网络和信息集成教育部重点实验室, 南京211189; 3东南大学中法生物医学信息研究中心, 南京210096
Author(s):
Bi Hui1;  Yang Guanyu1; 2; 3;  Tang Hui1 2; 3;  Shu Huazhong1; 2; 3
1School of Computer Science and Engineering, Southeast University, Nanjing 210096, China
2Key Laboratory of Computer Network and Information Integration of Ministry of Education, Southeast University, Nanjing 211189, China
3Centre de Recherche en Information Biomédicale Sino-Francais, Southeast University, Nanjing 210096, China
关键词:
超声图像分割;  Gabor特征;  局部二值模式;  k均值算法;  主动形状模型
Keywords:
ultrasound image segmentation;  Gabor features;  local binary pattern;  k-means algorithm;  active shape model
分类号:
TP391
DOI:
10.3969/j.issn.1001-0505.2017.05.007
摘要:
为了提高前列腺超声图像的分割精度,提出了一种基于改进主动形状模型的前列腺超声图像分割算法.首先,提取前列腺超声图像的特征集合,该特征集合由Gabor纹理特征和局部二值模式(LBP)特征组成.然后,通过利用k均值算法对提取的特征集合进行聚类分析,得到超声图像的聚类表示图.最后,在聚类表示图上应用ASM获取超声图像中前列腺的形状信息.结果表明,该算法可以准确地定位前列腺边界信息,与医生手动标记的前列腺轮廓相比,平均绝对距离仅为1.559 6 mm,戴斯相似度系数最高可达93.88%.利用超声图像的聚类表示图可以获得更加精确的前列腺轮廓信息,可用于海扶高聚焦超声(HIFU)手术中的精准导航.
Abstract:
To improve the segmentation accuracy of prostate ultrasound images, a prostate ultrasound image segmentation algorithm based on the improved active shape model(ASM)is proposed. First, the prostate ultrasound image feature set consising the Gabor features and the local binary pattern(LBP)features is extracted. Then, the cluster analysis of the feature set is carried out by using the k-means algorithm to obtain the clustering representation of the ultrasound image. Finally, based on the clustering representation, the shape information of the prostate is obtained by ASM. The results show that the proposed algorithm can precisely locate the prostate contour. Compared with the doctor delineation of the prostate, the mean absolute distance(MAD)is only 1.56 mm. The highest Dice similarity coefficient(DSC)reaches 93.88%. The segmentation based on the clustering representation of the ultrasound image can achieve more accuracy results and can be applied on precision navigation of high intensity focus ultrasound(HIFU)clinical operation.

参考文献/References:

[1] Amer Cancer Soc. Cancer facts and figures 2016 [EB/OL].(2017-01-31). http://www.cancer.org.
[2] Noble J A, Boukerroui D. Ultrasound image segmentation: A survey [J]. IEEE Transactions on Medical Imaging, 2006, 25(8): 987-1010. DOI:10.1109/tmi.2006.877092.
[3] Xie J, Jiang Y, Tsui H. Segmentation of kidney from ultrasound images based on texture and shape priors [J]. IEEE Transactions on Medical Imaging, 2005, 24(1): 45-57.
[4] Shen D, Zhan Y, Davatzikos C. Segmentation of prostate boundaries from ultrasound images using statistical shape model [J]. IEEE Transactions on Medical Imaging, 2003, 22(4): 539-551. DOI:10.1109/TMI.2003.809057.
[5] Gong L, Pathak S D, Haynor D R, et al. Parametric shape modeling using deformable superellipses for prostate segmentation [J]. IEEE Transactions on Medical Imaging, 2004, 23(3): 340-349. DOI:10.1109/tmi.2004.824237.
[6] Zhan Y, Shen D. Deformable segmentation of 3-D ultrasound prostate images using statistical texture matching method [J]. IEEE Transactions on Medical Imaging, 2006, 25(3): 256-272. DOI:10.1109/TMI.2005.862744.
[7] Michailovich O, Adam D. A high-resolution technique for ultrasound harmonic imaging using sparse representations in Gabor frames [J]. IEEE Transactions on Medical Imaging, 2002, 21(12): 1490-1503. DOI:10.1109/TMI.2002.806570.
[8] Ojala T, Pietikainen M, Maenpaa T. Multiresolution gray-scale and rotation invariant texture classification with local binary patterns[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(7): 971-987. DOI:10.1109/tpami.2002.1017623.
[9] van Ginneken B, Frangi A F, Staal J J, et al. Active shape model segmentation with optimal features[J]. IEEE Transactions on Medical Imaging, 2002, 21(8): 924-933. DOI:10.1109/TMI.2002.803121.
[10] Cootes T F, Taylor C J, Cooper D H, et al. Training models of shape from sets of examples [C]//British Machine Vision Conference. London, England, 1992: 9-18. DOI:10.5244/c.6.2.
[11] Cootes T F, Cootes C J. Active shape model search using local grey-level models: A quantitative evaluation [C]//British Machine Vision Conference. London, England, 1993: 639-648.
[12] Ladak H M, Mao F, Wang Y, et al. Prostate boundary segmentation from 2D ultrasound images [J]. Medical Physics, 2003, 27(8): 1648-1659. DOI: 10.1118/1.1286722.
[13] Cosío F A. Automatic initialization of an active shape model of the prostate [J]. Medical Image Analysis, 2008, 12(4): 469-483. DOI:10.1016/j.media.2008.02.001.
[14] Yan P, Xu S, Turkbey B, et al. Discrete deformable model guided by partial active shape model for TRUS image segmentation[J]. Transactions on Biomedical Engineering, 2010, 57(5): 1158-1166. DOI:10.1109/TBME.2009.2037491.
[15] Ghose S, Oliver A, Mitra J, et al. A supervised learning framework of statistical shape and probability priors for automatic prostate segmentation in ultrasound images [J]. Medical Image Analysis, 2013, 17(6): 587-600. DOI:10.1016/j.media.2013.04.001.

备注/Memo

备注/Memo:
收稿日期: 2017-01-09.
作者简介: 毕卉(1985—),女,博士生;舒华忠(联系人),男,博士,教授,博士生导师,shu.list@seu.edu.cn.
基金项目: 国家自然科学基金资助项目(31571001, 61201344, 61271312, 61401085, 81530060)、江苏省自然科学基金资助项目(BK2012329, BK2012743, BK20150647, DZXX-031, BY2014127-11).
引用本文: 毕卉,杨冠羽,唐慧,等.基于改进主动形状模型的前列腺超声图像分割算法[J].东南大学学报(自然科学版),2017,47(5):879-883. DOI:10.3969/j.issn.1001-0505.2017.05.007.
更新日期/Last Update: 2017-09-20